基于机器视觉的机械零件边缘像素补偿法  

Pixel Compensation Method of Machanical Part Edge Based on Machine Vision

作　　者：董芳凯 郑智贞[1] 袁少飞[1] 白云鑫 王宁[1] 

机构地区：[1]中北大学

出　　处：《工具技术》2017年第9期132-134,共3页Tool Engineering

摘　　要：机械零件图像边缘灰度的分布特征已成为影响机器视觉准确测量机械零件的重要因素之一。为了减小这种测量误差,提高测量精度,提出边缘像素补偿法来对机械零件图像边缘进行处理。通过分析视觉边缘检测原理,应用高精度标准块对相机进行标定。标定后对待测零件进行多次测量,试验结果表明:对图像信息进行边缘补偿后,其测量结果的相对误差可减小0.059%,边缘补偿算法可以有效提高测量精度。The distribution characteristics of edge image gray of mechanical parts have become one of the important factors that affect machine vision to accurately measure mechanical parts. In order to reduce the measurement error and improve the measurement accuracy, an edge pixel compensation method is proposed to deal with the edge of mechanical parts image. By analyzing the principle of vision edge detection, the camera is calibrated by high precision standard block. The re- sults showed that the relative error of the measured result decreases O. 059% after the edge compensation of image information. Edge compensation algorithm can effectively improve the measurement accuracy.

关 键 词：视觉检测 边缘补偿 图像处理 相机标定 

分 类 号：TG806[金属学及工艺—公差测量技术] TH741[机械工程—光学工程]